JP4505930B2 - Local lattice strain measuring method and manufacturing method of micro device using the same - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a local lattice strain measurement method that makes it possible to measure a lattice strain in a local region of a crystal substrate simply and with high accuracy using a focused electron beam, and a method for manufacturing a micro device using the same.
[0002]
[Prior art]
With the progress of high integration and miniaturization of integrated circuit devices and the like, it is known that the lattice strain locally generated in the crystal substrate during the element manufacturing process has a greater effect on element characteristics and reliability. In particular, in recent years, the STI structure has been used for element isolation of semiconductor devices, and the influence of local lattice distortion has become even stronger. Therefore, it is desired to put to practical use a method for measuring the lattice strain in the local region of the crystal substrate simply and with high accuracy.
[0003]
Microscopic Raman spectroscopy is generally known as a method for measuring lattice strain in a local region. However, since it is difficult to obtain a spatial resolution exceeding 1 μm with this method, a semiconductor device having a minimum dimension on the order of submicron is used. It cannot be an effective method for manufacturing.
[0004]
On the other hand, the transmission electron microscope can converge an electron beam down to the order of several nanometers. Therefore, it can be used as an effective means for evaluating local lattice strains on the order of submicrons. Line diffraction methods are known. In this method, the electron beam is converged and irradiated on the crystal substrate, and the diffraction pattern (higher order Laue zone) appearing on the transmission disk is compared with the pattern obtained by simulation to determine the electron beam acceleration voltage and the crystal substrate lattice. A constant is calculated.
[0005]
[Problems to be solved by the invention]
The convergent electron beam diffraction method can be applied to manufacturing process management of a miniaturized semiconductor device in terms of spatial resolution and measurement accuracy of lattice distortion. However, in the conventional focused electron diffraction method, the procedure for matching the actually obtained pattern with the pattern obtained by the simulation is complicated, so that there is a problem that it is practically difficult to use it for manufacturing process management of semiconductor devices. there were.
[0006]
Therefore, the present invention provides a simple method for measuring the lattice strain of a crystal substrate with high spatial resolution, and improves device characteristics and reliability by manufacturing minute devices using this method. The purpose is to let you.
[0007]
[Means for Solving the Problems]
The solution to the above problem is to converge the electron beam to irradiate a local region of the crystal substrate, and among the transmitted electron beam and diffracted electron beam that appear on the screen, a set of Kikuchi lines consisting of parallel defect lines and excess lines. Measuring the distance between the deficient line and the excess line, and measuring the lattice strain of the local region based on the result,
Alternatively, a higher-order higher-order Kikuchi line is used, and the crystal substrate is inclined at an angle of 15 degrees or less about a direction perpendicular to the Kikuchi line, and the missing line of the Kikuchi line set and the Measuring the distance between the excess lines and measuring the local lattice strain,
Alternatively, it is achieved by a method for manufacturing a micro device characterized in that an element is formed in a crystal base material having a small lattice strain selected based on a measurement result obtained by the local lattice strain measurement method.
[0008]
FIG. 1 is a diagram illustrating the principle of the present invention. As schematically shown in the figure, when the convergent electron beam 1 is incident on the surface of the thinned crystal substrate 2 with a slight inclination from the direction perpendicular to the surface, a linear shape consisting of a deficient line and an excess line parallel to each other. No. 5 and 6 of Kikuchi Line appear on the screen 4. As seen in the figure, one of the pairs 5 and 6 of the Kikuchi line is an electron beam diffracted by a lattice plane (not shown) passing through the P point and transmitted through the crystal substrate 2 as it is, and a lattice passing through the P point. This occurs when the electron beam that has been diffracted by the surface and further diffracted by the lattice plane 3 passing through the R point and transmitted through the crystal substrate 2 interferes, and the other is diffracted by the lattice plane passing through the P point and is directly crystallized. This is caused by interference between the electron beam transmitted through the substrate 2 and the electron beam diffracted by the lattice plane 3 passing through the point P and then diffracted by the lattice plane 3 passing through the point Q and transmitted through the crystal substrate 2.
[0009]
Then, it is possible to measure the distance between the missing line and the excess line forming the set of the Kikuchi line, and measure the change in the lattice constant from the measurement result. Since the distance between pairs of Kikuchi lines increases as higher-order Kikuchi lines are used, the measurement accuracy can be increased. According to the inventor's experiment, it was found that a measurement accuracy of 1 × 10 ⁻³ or more can be obtained by using a higher-order Kikuchi line of the third order or higher.
[0010]
In FIG. 1, a so-called Kikuchi band appears on the screen 4 when the convergent electron beam 1 is perpendicularly incident on the crystal substrate 2. This hinders the measurement of the distance between the pair 5 and 6 of the Kikuchi line that appears corresponding to the lattice plane 3 to be measured, but the strength of the Kikuchi band becomes weaker as it becomes higher. The influence can be avoided by using a high-order Kikuchi line.
[0011]
In general, many Kikuchi lines that are not parallel to the Kikuchi line used for measurement of a specific lattice plane appear on the screen 4, but the crystal base material is slightly inclined in a specific direction according to the measured lattice plane, By using higher-order Kikuchi lines higher than the third order, the influence of these Kikuchi lines can be avoided. If the inclination angle is increased more than necessary, the strength of the Kikuchi line will decrease, making measurement difficult. The inventors have experimentally found that an appropriate tilt angle range is within ± 15 degrees.
[0012]
Further, the inventors have confirmed through experiments that the electron beam must be converged to 1 mrad or more in order to produce a clear diffraction image on the screen in the above measurement.
[0013]
In addition, the local lattice strain distribution of the crystal base material is measured before and after the specific process in the element manufacturing process, and the influence of the specific process on the crystal base material is evaluated. It is possible to improve the characteristics and reliability of the element by forming the element in a selected crystal base material having a small lattice strain.
[0014]
DETAILED DESCRIPTION OF THE INVENTION
In manufacturing a 0.18 μm CMOS device, an STI structure is used as an element isolation structure of a Si wafer. Therefore, an example in which process management is performed by applying the local lattice strain measurement method according to the present invention to the STI formation process for the Si wafer and measuring the lattice strain during the process will be described below.
[0015]
2A to 2E show a process of forming an STI structure on a Si wafer having a (100) plane. First, as shown in FIG. 1A, the surface of the Si wafer 10 is thermally oxidized to deposit a thermal oxide film 11 having a thickness of 3 nm, and then nitrided to a thickness of 100 nm by using the CVD method. A film 12 is deposited. Further, subsequently, a hard mask oxide film 13 having a thickness of 300 nm is deposited using the CVD method to form a photoresist pattern 14 for forming an STI region.
[0016]
Next, as shown in FIG. 2B, the trench 15 is formed by anisotropic dry etching using the resist pattern 14 as a mask. Then, as shown in FIG. 3C, an oxide film 16 is deposited by CVD to fill the trench 15 inside.
[0017]
Next, as shown in FIG. 4D, the surface of the Si wafer 10 is planarized by CMP using the nitride film 12 as a stopper. Then, the oxide film 16 is densified by high-temperature heat treatment in a nitrogen atmosphere. Finally, as shown in FIG. 4E, the thermal oxide film 11 and the nitride film 12 are removed by etching, whereby an STI structure is formed on the Si wafer 10.
[0018]
Next, the result of applying the local lattice strain measurement method according to the present invention to the management of the STI formation process will be described. In the present embodiment, one of the Si wafers that have undergone the process described in FIG. 2C, that is, the Si wafer that has been filled in the trench 15 with the oxide film 16 and that has not been subjected to the heat treatment, is extracted for the first time. Measure the lattice strain.
[0019]
Further, three Si wafers that have been subjected to the process described in FIG. 2D, that is, the Si wafer after the surface of the Si wafer is planarized by the CMP method and heat-treated are extracted to each Si wafer. On the other hand, heat treatment is performed under the following three different conditions, and the second lattice strain measurement is performed for each of the Si wafers subjected to these heat treatments. The heat treatment conditions were as follows: (1) 950 ° C., 30 minutes, (2) 1000 ° C., 30 minutes, (3) 1050 ° C., 30 minutes in a nitrogen atmosphere. Then, by comparing the first measurement result and the second measurement result, the change in lattice strain of the Si wafer before and after the heat treatment was examined, and the optimum heat treatment conditions were set.
[0020]
FIG. 3 shows a schematic diagram and a diffraction image for explaining a local lattice strain measuring method for the Si wafer. The sample used for the measurement was prepared by the same method as that for a normal transmission electron microscope observation sample using mechanical polishing and ion milling of a Si wafer. In the figure, reference numeral 20 denotes a measurement sample prepared by slicing a (100) surface Si wafer in a direction perpendicular to the surface, and the plane orientation of the cross section of the sample 20 is [1,- 1, 0]. An STI 23 is formed on the sample surface 21, and the thickness of the sample is 0.3 μm.
[0021]
As shown in FIG. 3, when the convergent electron beam 22 is irradiated with being slightly inclined with respect to the cross section of the sample 20, a defect line and an excess line are formed on the fluorescent plate 24 arranged on the opposite side of the convergent electron beam 22 across the sample 20. A pair of Kikuchi Lines will appear. Here, the 0012 deficit line and the 0012 excess line are used for the measurement of the plane spacing of the (001) lattice plane parallel to the sample surface 21, and the plane spacing of the (110) lattice plane perpendicular to the sample surface 21 is used. Used an 880 deficit line and an 880 excess line. FIG. 3 shows a diffraction image observed when the electron beam 22 converged to 10 mrad is irradiated. The inventors have confirmed that a clear diffraction image can be obtained by using an electron beam converged to 1 mrad or more.
[0022]
FIGS. 4A to 4C show a procedure for measuring the plane spacing of the (001) lattice plane parallel to the sample surface 21. FIG.
[0023]
First, in order to set conditions for allowing the convergent electron beam 22 to be incident perpendicularly to the cross section of the sample 20, the orientation of the converged electron beam 22 is adjusted so that the Kikuchi band appears symmetrically on the fluorescent plate 24. FIG. 4A shows a state in which the 004 Kikuchi band appears at a symmetrical position on the screen, and a disk made of a transmission electron beam appears at the center. When the sample is tilted from this state in a direction parallel to the sample surface 21, as shown in FIG. 4B, a set of tertiary Kikuchi lines, that is, a 0012 deficit line (indicated by a dotted line). And the 0012 excess line (shown as a solid line) are excited. This figure shows the state when the sample 20 is tilted until the 0012 defect line crosses almost the center of the transmission disk, and the tilt angle at this time is about 3.5 degrees.
[0024]
Next, set the orientation so that other Kikuchi lines do not interfere with the 0012 Kikuchi line by tilting the sample in the direction perpendicular to the axis perpendicular to the Kikuchi line without changing the arrangement of the 0012 missing line visible in the transmission disk. did. As can be seen in FIG. 4C, a good diffraction image could be obtained by tilting about 10 degrees.
[0025]
Next, FIGS. 5A to 5C show a procedure for measuring the spacing of the (110) lattice plane perpendicular to the sample surface 21. FIG.
[0026]
First, the direction in which the focused electron beam is incident perpendicularly to the cross section of the sample 20 is set. This can be set by adjusting the incident direction of the convergent electron beam so that the Kikuchi band appears symmetrically on the fluorescent plate 24 as in FIG. The state at this time is shown in FIG. Next, the sample 20 is slightly tilted from the center of incidence of the convergent electron beam 22 while maintaining an orientation parallel to the sample surface 21, and as shown in FIG. 5B, an 880 defect line (appears as a dotted line). And the 880 excess line (indicated by the solid line) was excited. This is done by inclining in the direction parallel to the sample surface while looking at the Kikuchi line from the state shown in FIG. 5A, and setting so that the 880 defect line appears in the direction crossing almost the center of the transmission disk. can get. The inclination angle at this time was about 3.2 degrees.
[0027]
Next, the sample 20 was tilted about the axis perpendicular to the Kikuchi line so as not to change the arrangement of the 880 missing lines visible in the transmission disk, and the direction in which the other Kikuchi lines did not interfere with the 880 Kikuchi lines was set. . In the figure, a favorable diffraction image could be obtained by tilting about 7 degrees.
[0028]
For distance measurement between Kikuchi line pairs that appeared on the screen, the above-mentioned fluorescent plate was used as a screen, and the Kikuchi lines that appeared on this fluorescent plate were baked on a negative film for a microscopic electron microscope. Can be used to obtain the distance between the Kikuchi lines. According to the inventor's experiment, it was possible to easily obtain a measurement accuracy of about 1 × 10 ⁻³ by using a negative film having a size of 5.9 × 8.2 cm and a resolution of 10 μm or more.
[0029]
In order to further improve the measurement accuracy, the image obtained with the above negative film is printed on another negative film at a predetermined magnification and then read with a scanner. The image obtained with the negative film is captured with a scanner. In some cases, a method using a magnifying scanner with a lens, a method of capturing an image obtained with a negative film with a CCD camera, or the like can be used. Measurement accuracy of 1 × 10 ⁻⁴ or more can be obtained by these methods.
[0030]
Moreover, it can replace with a negative film and can also measure using an imaging plate. For example, it was possible to obtain a measurement accuracy of 1 × 10 ⁻³ or more by reading with a distance of 5 cm between the pairs of Kikuchi lines on an imaging plate having a resolution of about 25 μm.
[0031]
Furthermore, the same measurement accuracy can be obtained even if a CCD camera having a resolution of about 25 μm is used instead of the negative film.
[0032]
In the above-described embodiment, by using the 0012 Kikuchi line and the 880 Kikuchi line for the Si wafer having the (100) plane, the intervals between the lattice planes parallel to and perpendicular to the surface of the Si wafer are obtained. Although the method has been described, for a Si wafer having a (111) plane, a lattice perpendicular to the lattice plane parallel to the surface of the Si wafer is obtained by exciting the 12 −6 −6 Kikuchi and 0 −88 Kikuchi lines. The distance between the surfaces can be determined.
[0033]
FIGS. 6A and 6B show the lattice strain distribution obtained by the above-described measurement for the Si wafer having the STI structure. The intersection of the arrows indicates the center of the local region where the measurement is performed, and the arrows Represents the magnitude of the lattice strain. FIG. 4A shows the first measurement result before the heat treatment described above. It was found that the lattice strain was the maximum near the edge of the trench, and the magnitude was 7 × 10 ⁻⁴ .
[0034]
FIG. 4B shows the second measurement result after the heat treatment using the heat treatment condition (2) described above, and the maximum value of the lattice strain near the edge of the trench is 7 × 10 ^−4. It was found to drop to 2 × 10 ⁻⁴ . As a result of performing the same measurement on the lattice strain when the heat treatment conditions (1) and (3) described above were used, it was found that the maximum values were 5 × 10 ⁻⁴ and 4 × 10 ⁻⁴ , respectively. This is because the thermal recovery is insufficient under the heat treatment condition (1) and the degree of relaxation of the lattice strain is weak, and under the heat treatment condition (3), the oxidation of the Si wafer surface is promoted by the excessive heat treatment and the lattice strain is reversed. This indicates that the heat treatment condition (2) is optimal for relaxing the lattice strain. This can be confirmed from the experimental results described below.
[0035]
FIG. 7 shows the experimental results showing the dependence of the leakage current of the gate oxide film on the heat treatment conditions in a MOS device fabricated using the Si wafer that has undergone the STI formation step. An electric field applied to a gate oxide film after applying an electrical stress by forcing a current of 0.05 A / cm ² for 5 seconds to a MOS device having a gate oxide film with an area of 0.1 mm ² This is a measurement of the relationship with the leakage current. Considering that the leakage current resistance increases as the electric field at which the leakage current begins to flow increases, this figure shows that the resistance against leakage current is maximized when the heat treatment condition (2) is used. This also agrees with the measurement result of the lattice strain described in 1. From this, it becomes possible to evaluate the leakage current resistance based on the lattice strain measurement result.
[0036]
From the above, process processing conditions can be optimized by measuring the lattice strain during the manufacturing process, and device characteristics can be improved by eliminating Si wafers having a lattice strain of a predetermined value or more during the process. It has become possible to improve the device reliability.
[0037]
The above embodiment describes an embodiment in which the present invention is applied to a semiconductor device. However, even if it is applied to other devices, the same effect can be obtained by the same action. For example, the LCD ( The present invention can be applied to a device that needs to form minute elements in a crystal substrate, such as a liquid crystal display device, a PDP (plasma display) device, a magnetic disk head, and a printer head.
[0038]
【The invention's effect】
As described above, in the present invention, by measuring the distance between the pairs of Kikuchi lines that appear on the screen, the lattice strain can be easily measured with a high spatial resolution in nm units and with an accuracy of about 1 × 10 ^−4. In addition, it is possible to improve the device characteristics and reliability by applying this method to the management of the manufacturing process of a micro device such as an LSI of submicron order.
[Brief description of the drawings]
FIG. 1 is a diagram illustrating the principle of the present invention. FIG. 2 is a cross-sectional view showing an STI formation process. FIG. 3 is a schematic diagram illustrating an embodiment of the present invention and an observed diffraction image. Figure to show (the 1)
FIG. 5 is a diagram showing a procedure for measuring lattice strain (No. 2)
FIG. 6 is a cross-sectional view showing the distribution of lattice strain. FIG. 7 is a view showing the dependence of leakage current on heat treatment conditions.
1, 22 Convergent electron beam 13 Oxide film for hard mask 2, 20 Sample 14 Resist pattern 3 Lattice surface 15 Trench 4 Screen 16 Oxide film 5, 6 Kikuchi Line 21 Sample surface
10 Si wafer 23 STI
11 Thermal oxide film 24 Fluorescent screen
12 Nitride film

Claims (3)

And converging the electron beam or the 1mrad irradiating a localized area of the crystal base, of the transmitted electron beam and the diffracted electron beam appearing on a screen, a set of 該欠loss of Kikuchi lines consisting of parallel-deficient lines and excessive line A local lattice strain measuring method, comprising: measuring a distance between a line and the excess line; and measuring a lattice strain of the local region based on the result. Using a higher-order Kikuchi line of 3rd order or more, and tilting the crystal base material at an angle of 15 degrees or less about a direction perpendicular to the Kikuchi line, the missing line and the excess line of the Kikuchi line set The local lattice strain measuring method according to claim 1 , wherein a distance between the two is measured.   A method for manufacturing a microscopic device, wherein an element is formed in a crystal base material having a small lattice strain selected based on a measurement result obtained by the local lattice strain measuring method according to claim 1.

Images